Multi-Cluster and Multi-Tenancy
You'll learn how service meshes connect services across different clusters or clouds while enforcing security and traffic policies.
What This Means
In real-world deployments, you often run services across multiple Kubernetes clusters (for redundancy, regional proximity, or compliance) or need to isolate tenants (customers) sharing the same infrastructure. A service mesh can extend its control across these boundaries, making cross-cluster communication transparent while enforcing the same policies you already learned—mTLS, authorization, traffic routing—everywhere.
Why It Matters
Imagine you have:
- Multiple clusters in different AWS regions for disaster recovery
- Multi-tenancy where CustomerA and CustomerB share infrastructure but must never access each other's data
Without a mesh, you'd manually configure VPNs, firewall rules, and certificates for each cross-cluster connection. With a mesh supporting multi-cluster:
- Services discover and call each other across clusters as if local
- mTLS works seamlessly between clusters (certificates issued centrally)
- Traffic policies (retries, circuit breaking) apply uniformly
- Tenant isolation is enforced through namespaces and authorization rules
The Multi-Cluster Pattern
The service mesh control plane either replicates to each cluster (federated model) or runs centrally with agents in each cluster. Services in Cluster A can call payment-service in Cluster B transparently—the sidecar proxy routes it correctly, verifying identity via mTLS.
Multi-Tenancy Security
For tenants, the mesh enforces namespace-level isolation: Tenant A's services can't send traffic to Tenant B's namespace, even though both run on the same physical cluster. Authorization policies define these boundaries.
Key Takeaway:
Multi-cluster and multi-tenancy features let service meshes extend secure, policy-driven communication across physical boundaries—making distributed infrastructure feel like one logical fabric.